Systematic temperature error in RCARS diagnostics from improper Raman linewidths

The coherent anti-Stokes Raman spectroscopy (CARS) stands as the standard for thermometry and major species detection and quantification in gas phase and combustion diagnostics. In recent years the database of empirical S-branch Raman linewidth for a variety of gases, gas mixtures and temperatures for rotational CARS spectroscopy has been significantly expanded. The Raman linewidths are of utmost importance for accurate extraction of thermodynamic information from the spectral information. However, unavailable correct linewidth data for rotational CARS evaluations are regularly substituted by approximated data for example by omission of the proper collisional environment or by using the vibrational Q-branch linewidth instead of the rotational S-branch linewidths. The resulting systematic errors by using incorrect linewidths have only been studied for a few selected cases which hint at a significant temperature error of up to 9 %. In this work we show a clear picture of the temperature and concentration errors resulting from incorrect linewidth data by evaluating the influence of S- vs. Q-branch linewidths in pure oxygen and self-broadening vs. accurate collisional environment in air and nitrogen-water vapor mixtures. The data show that the evaluated temperature is systematically too high by using Q-branch instead of S-branch linewidths in oxygen and nitrogen thermometry and a strong influence of the collisional environment on the temperature and species concentration determination.